H. S. Washington — Rhyolites of Lipari. 457 



show marked fluidal arrangement. No microphenocrysts 

 are to be seen. 



The analysis is given in Table IV. 



TABLE IV. 



1 2 



SiO, 76.56 76.73 



AloO 13.47 12.24 



Fe,0 0.20 1.38 



Feb 1.01 0.76 



MgO 0.08 0.22 



CaO 1.47 1.46 



Na,0 3.76 3.32 



K,6 3.51 4.17 



H„0+ 0.27 ) 



V 0.49 



H,0— 0.22 ) 



TiO. 0.10 n.d. 



P,0, n.d. n.d. 



SO, n.d. n.d. 



CI n.d. n.d. 



MnO 0.06 n.d. 



100.71 100.77 



1. — Rhyolite obsidian, Bombarda, Milos. 1.3 (4) .2.3 (4). Washington, an- 

 alyst. 



2. — Rhyolite obsidian, Jali, Nisyros. 1.3 (4) .2.3. Martelli, analyst. A. 

 Martelli, Gruppo di Nisiro, p. 69, Rome, 1917. 



As compared with the Lipari obsidians, this shows 

 somewhat higher silica, as well as higher lime, somewhat 

 lower alkalies, with a rather higher proportion of soda. 

 These differences are clearly and concisely expressed by 

 the symbols according to the quantitative system. It will 

 be noted that in this highly vitreous obsidian the ferrous 

 oxide is greatly in excess of the ferric, just as we have 

 seen to be the case elsewhere. 



My specimens* of the other Milos rocks are not available 

 at present, so I am unable to make an analysis now of 

 the noncrystalline varieties. Study of their thin sections 

 indicates, however, that most of those in my possession 

 are rhyolites, with orthoclase or sodi-orthoclase as 

 phenocrysts and groundmass laths, and more or less 

 quartz. But a little plagioclase seems to be also gener- 

 ally present, so that these rocks may be considered as 

 toscanites, rather than purely alkalic rhyolites. This 

 occurrence of plagioclase is in agreement with Ehren- 

 burg's descriptions of many andesites on Milos, though 

 few of these are represented in my small collection of 



